Method ReadBatchWithDictionary

cpp/src/parquet/column_reader.cc:1051–1096

template <typename DType>
int64_t TypedColumnReaderImpl<DType>::ReadBatchWithDictionary(
    int64_t batch_size, int16_t* def_levels, int16_t* rep_levels, int32_t* indices,
    int64_t* indices_read, const T** dict, int32_t* dict_len) {
  bool has_dict_output = dict != nullptr && dict_len != nullptr;
  // Similar logic as ReadValues to get pages.
  if (!HasNext()) {
    *indices_read = 0;
    if (has_dict_output) {
      *dict = nullptr;
      *dict_len = 0;
    }
    return 0;
  }

  // Verify the current data page is dictionary encoded.
  if (this->current_encoding_ != Encoding::RLE_DICTIONARY) {
    std::stringstream ss;
    ss << "Data page is not dictionary encoded. Encoding: "
       << EncodingToString(this->current_encoding_);
    throw ParquetException(ss.str());
  }

  // Get dictionary pointer and length.
  if (has_dict_output) {
    GetDictionary(dict, dict_len);
  }

  // Similar logic as ReadValues to get def levels and rep levels.
  int64_t num_def_levels = 0;
  int64_t indices_to_read = 0;
  ReadLevels(batch_size, def_levels, rep_levels, &num_def_levels, &indices_to_read);

  // Read dictionary indices.
  *indices_read = ReadDictionaryIndices(indices_to_read, indices);
  int64_t total_indices = std::max<int64_t>(num_def_levels, *indices_read);
  // Some callers use a batch size of 0 just to get the dictionary.
  int64_t expected_values = std::min(batch_size, this->available_values_current_page());
  if (total_indices == 0 && expected_values > 0) {
    std::stringstream ss;
    ss << "Read 0 values, expected " << expected_values;
    ParquetException::EofException(ss.str());
  }
  this->ConsumeBufferedValues(total_indices);

  return total_indices;
}

template <typename DType>
int64_t TypedColumnReaderImpl<DType>::ReadBatch(int64_t batch_size, int16_t* def_levels,